A Crack in the Edge of the World

by Simon Winchester

  And the Mississippi, mighty as it was—in places more than half a mile across—began to behave in ways that defied credulity. After one particularly dramatic sequence of shaking, it developed violent overfalls and began to flow backward, its waters surging in borelike shudders up toward the mouth of the Ohio, and spuming as they battled against the furious downstream torrent. A Scotsman named John Bradbury, sent over from Liverpool to catalog the botanical life of the American West, happened to be on the river during one of these episodes. His boat, he later wrote, suddenly and for no obvious reason, lurched upward, so rudely that everyone imagined an imminent capsize. Bradbury was knocked off his feet, but, when he grabbed hold of the railing and tried to stand, he saw that the river was covered with foam, and thick with broken branches and whole trees, the entire surface whipped up into a lather as if during a fearful storm—and yet there was no wind, no lightning, no rain. “The noise was inconceivably loud and terrific,” he said, although he could “distinctly hear the crash of falling trees, and the screaming of the wild fowl on the river … all nature was in a state of dissolution.” The riverbanks were caving in with enormous crashes, jagged chasms were being torn in the low cliffs, and dozens of boats—all empty and forlorn—were drifting past in the frenzy, evidence, so Bradbury supposed, of the deaths of all those caught up in the terror.

  The botanist, who was clearly well trained in scientific observation, was able to count twenty-seven distinct additional shocks as the night wore on, while his party’s little cutter was hove to, waiting for an end to whatever was happening. In time he noticed a distinct pattern: First there was a sudden sound, usually from “a little northwards of east,” and then there was a shock; and once that was over, the sound disappeared, quietening itself to a whisper as it did so, in what seemed to Bradbury was a generally westerly direction. In time it diminished entirely, and Bradbury’s party made its way downriver—eventually reaching Natchez in the new year, 1812. There he saw the arrival of the steam-driven stern-wheeler New Orleans, the first steam-powered vessel to navigate the Mississippi: It had passed through New Madrid while the water was foaming and the surviving inhabitants were wailing in terror and, he supposed, demanding to be picked up. But, he reported, the captain had decided not to stop—partly because he was afraid of berthing while the earthquake was going on, and partly because he soon realized from the wailing of the town’s inhabitants that they were as terrified by his mighty new vessel, belching smoke and making its fearsome engine noise, as they were by what was happening beneath their feet.

  The number of events in the sequence was truly prodigious. Jared Brooks, of Louisville, Kentucky, counted no fewer than 1,874 separate earthshaking episodes in and around New Madrid over the next few weeks. Shocks like the first enormous one of December 16 occurred two more times—once on January 23 and again on February 7, this last being the mightiest of all. And then the world fell silent again. New Madrid rebuilt itself to the extent that it ever would, and quietly resumed its place as an otherwise forgettable little town in the middle of an otherwise most forgettable corner of the American Midwest.

  Except that it remains a very seismically active place today, a place where the needles on seismographs mounted in the tiny city museum down by the levee wave furiously all the time, like antennae of some excitable insects. A number of faults have recently been identified, running deep and hidden below the alluvial soil that today nourishes the cotton, corn, and soybeans up above. Two of these faults trend from the northeast to the southwest, just as the Appalachian Mountains, now hundreds of miles away back east, do also; and the other two run at ninety degrees to them, appearing on the map as lines of cross-hatching.

  A map that shows the estimated epicenters of the thousands of earthquakes that occurred in 1811, and of the tens of thousands that have caused the land here to chatter away noisily ever since, follows the traces exactly. A line of black dots and other symbols of varying size and shape runs from the hamlet of Marked Tree in Arkansas up to the tiny town of Vienna, Illinois, northeast to southwest, illustrating occurrences of one of the sets of quakes; and another line seems to form a crosspiece over the first one, running as it does between Fisk, Missouri, and Trenton, Tennessee, and sporting another array of dots, triangles, and circles, revealing how the quakes followed that fault line, too.

  The two lines intersect in New Madrid, at a place shown on the seismic map as a wildly confusing maze of dots, dashes, lines, stars, circles, and other coded markings. These indicate that this is perhaps the most active, and in time possibly the most potentially deadly, location in all of America. The relatively new science of paleoseismology—the looking for evidence of early earthquakes—tells us that this intersection has not only been plagued by small and occasional quakes in the past, but also suffered from sequences of truly enormous earthquakes. It seems to have experienced a major sequence around A.D. 900, and again around A.D. 1500. If these are taken into consideration with the 1811 event, it now looks as though New Madrid is being visited by chaos and lethality every three or four centuries—with the corollary that the end of the twenty-first century, perhaps, will be the time for the intersection of the fault lines to act up again.

  If it does so, matters will probably be infinitely more terrible. Not so much because of New Madrid’s future as a hub of commerce and population—such a thing will probably never happen. It will be terrible because the epicenter exists, poised to rupture, at the still center of a full one million square miles of American territory that has now within its borders cities, factories, waterways, military bases, suburbs, and slums that would all be seriously, devastatingly affected by a series of quakes of the same size as those that took place in 1811. It is likely that cities will be bigger, with populations more densely settled, roads and railway lines more important, factories more sophisticated, and—unless lessons are learned quickly—all settled and built with only a limited awareness of what the earth below may be readying itself to do.

  When the faults ruptured in 1811, the western half of the earthquake’s penumbra was to all intents and purposes uninhabited, with the other half peopled by pioneers and farmers and men and women who had made for themselves only modest dwellings in equally modest and still-growing towns. Now, however, from Chicago to Memphis,* from Oklahoma City to Nashville, there are giant conurbations crammed with people and with valuable buildings and priceless businesses. Few of the inhabitants spend much of their time today remembering that they have chosen or chanced to be in a notoriously dangerous earthquake zone. And if they ever do, then they generally reassure themselves by saying, without any statistical justification for doing so, that the winter of 1811 was when the region got its pent-up seismicity out of its system, and all is likely to be peaceful from now on.

  I HAD ONE MORE PLACE to visit before I entered the American West proper, before I began to travel into that part of the country that came under the direct influence of the specific geological peculiarities that generated and triggered—or, as geology likes to put it, were the ultimate and proximate causes of—the San Francisco Earthquake. My next immediate destination lay just a few hundred miles farther on from Missouri, at a general store in a tiny Oklahoma village with the distinctly unmemorable name of Meers.

  It had been named after a prospector, but nothing of substance had ever been found anywhere near it, by Mr. Meers or by anyone else. Back in the 1890s there had been all kinds of rumors of gold being found in the Wichita Mountains, including one tale of a housewife finding a nugget as big as a piece of buckshot in the craw of a fowl she was preparing for Christmas dinner. Such stories were often part of the western myth, tales circulated to encourage settlers to come out from back east. And some gullible, greedy or bored adventurers did indeed go there and build a clutch of shacks beside a creek. For a while their hopes were kept alive by town hucksters who had brought in some real gold miners from Colorado and salted a local shaft with imported nuggets. There was something of an Oklahoma gold rush, but then the mines
petered out, or were just abandoned as having been worthless all along.

  Before long such townsfolk as remained closed the mines, bought longhorn cattle from dealers in Texas, and began to ranch. Cattle ranching remains the staple of Meers today, though there is a lake, and a wildlife reserve with herds of buffalo, elk, and deer. The place positively hums at weekends when artillerymen from the huge base nearby at Fort Sill stop by to play. The first man I met at the Meers general store was an army colonel just back from serving two years in Bosnia. He was training on a new kind of howitzer and was expecting to be sent off to Iraq any day. Remote though the town may be, it has managed to forge some kinds of links with the world beyond.*

  This was amply displayed in 1985, when Meers began to excite, if only in a modest way, a keen interest in the worldwide community of seismologists. Teams of surveyors suddenly became fixated upon a strange escarpment that had apparently popped out of nowhere close by the town, and that ran for fifteen miles or so to the north, then ended as abruptly as it had begun. They couldn’t find any good reason for its existence—but once it was examined in detail, it was discovered that it marked the trace of a very distinct fault line. Moreover, it was a fault line that showed evidence of having moved, swiftly and violently, at least twice since it had been created. Two big earthquakes had apparently hit Meers—or rather, where Meers would eventually be, the more recent of the two quakes having taken place just over a thousand years ago. Exactly why this might be, no one at the time could tell.

  And so the Geological Survey of Oklahoma promptly put one of the mine shafts, abandoned for three-quarters of a century, to good use again. Technicians placed an array of seismometers deep inside the mine, then ran a coaxial cable up to a seismograph, and placed this in the only convenient, safe, and constantly monitored location they could find at the time, which was the Meers general store. It sits there still today, close by the main entrance, glanced at only occasionally by the cowboys coming in for their Meers Burgers and shakes and bottles of beer. The instrument’s recording drum turns slowly behind a protective plastic cover, its dials flicker, and its celebrity status is underscored by a vanity wall of yellowed press clippings that tell of its importance to the seismic world, and of the sensitivity of this particular machine. Earthquakes that happened on Diego Garcia, more than 10,000 miles away, were recorded on the drum at Meers, it says on the wall; but also, lest any visiting artilleryman try to show off, the guns firing out on the butts twenty miles away at Fort Sill never seem to cause the machine to record the slightest sympathetic tremor.

  So far the seismograph has recorded only one earthquake of any size that was definitely caused by movement on the Meers Fault—and since monitoring what everyone agreed was an alarmingly large and unpredictable fault was the purpose of the machine, some might regard this as a problem. This one movement was very small and took place in April 1997: since then, total seismic silence. Which makes seismologists all the more curious: Just when, the geological community wants to know, is the fault going to rupture?

  It is a classic of its kind. It is primed, tense, filled with stress, and ready to pop at any moment. But it seems to want to do so only every millennium or two—alarming for those who live around Meers (though in truth no one seems to mind one whit; and the owners of the store appreciate the curious travelers who drop by to inspect the machine), and alarming, too, because there are perhaps many thousands of faults just like those at Meers waiting to break and cause who knows how much mayhem, at unpredictable times of their own choosing.

  Knowing a little about Meers makes one far less confident, far more skeptical, about the newfangled science of earthquake prediction. The notion that we might one day be able to forecast quakes is a quest that ultimately motivates all seismic research—or the funding for it, at least—even if few seismologists care to admit it. And the sheer capriciousness of deeply hidden fault systems like those in Oklahoma serves mainly to remind researchers that it will be a long, long time before a predictive technique can be devised that will offer an anxious general public any useful degree of certitude. For although earthquakes do tend to happen in seismically obvious places—like San Francisco and Sumatra—they do also take place in the less predictable places, just out of seismic orneriness.

  The general store in Meers, Oklahoma, features enormous hamburgers in addition to a seismograph, which sits, ticking away imperturbably, beside one of the refrigerators.

  HERE IN THE MIDDLE of the North American Plate, far from the unstable edges where all the action takes place, are four locations where earthquakes simply should not happen but do: the Ramapo Fault in New York; the Summerville-Charleston axis in South Carolina; the New Madrid and Wabash Valley nexus in Missouri and its neighbor states; and the Meers Fault in Oklahoma. All of these are classified as intraplate events—they occur within the stable-sounding centers of tectonic plates. There are similar examples from around the world: Ungava in western Canada, Christmas Day 1989; Newcastle, Australia, in 1989, three days later; Killari, India, 1993.

  And in America, too, there have been yet others. Some, such as the huge quakes of central New Hampshire and Cape Ann, Massachusetts, are only dimly recorded and only in the geological record, since they happened long before there were Americans on hand to write accounts. More recently, there are records from sudden single events that puzzle and perplex to this day: the Grand Banks Earthquake of 1929 (which broke submarine cables and spawned waves that killed dozens in the outports of Newfoundland); and three still-notorious New York State events: Massena in 1944, Goodnow in 1983, Ardsley in 1985. They are sporadic, intermittent, unpredictable—and as a consequence they are much more dangerous to those living innocently near them than are the more generally expected earthquakes of California, where the cities make sure that buildings are made superstrong and the public is kept endlessly aware.

  The specific causes of these intraplate earthquakes are various. The proximate cause of events at Meers, for example, is explained away as the existence of, and occasional movement along, one very particular fault. The subterranean machine that causes the New Madrid quakes is driven by the reaction that occurs when two very different kinds of faults intersect and collide with each other, right under the town.

  The ultimate cause of events like these is more interesting, however—as well as being disarmingly simple.

  The best theory that anyone has these days for the underlying cause of all intraplate earthquakes is that they represent the relief of stresses built up eons ago, when the mountains or valleys or areas of basin and range were themselves being created. It is eminently reasonable to suppose that the plates themselves, being buckled and elevated and twisted and compressed at their edges, were subject to stress far away from where the buckling and elevation most dramatically occurred.

  A piecrust, as it is heated in the oven, will bow upward in the middle, where the fruit bulges and expands in the heat; it will also buckle up at the edges, where the heat conducted through the pan becomes most intense. In the area between the bow and the buckle, the crust, the second it emerges from the oven, lacks the nearly smooth and unblemished appearance it had when first placed inside: It now has myriad cracks and crevices, bulges and breaks. All of this may add to the pie’s charm, but it also suggests something of the stresses that built up in the crust as it was being cooked. Once the pie leaves the oven, and the temperature and pressure begin to ease, the cracks change and the aspect of the surface alters: The cracks start to widen or narrow, the surface perhaps tries to revert to its original unblemished and undistorted self, and the stresses that mounted during the cooking process are generally relieved—until a kind of stasis is achieved and the pie becomes what it will be, until it is eaten.

  And that, more or less, is what has happened and is still happening in the stable-sounding central parts of all tectonic plates. Almost all of the events just mentioned—Meers, Cape Ann, the Grand Banks, Ardsley—seem to have resulted from the relief of the kind of stress patterns that
are illustrated by the cooling piecrust.

  But New Madrid, as it happens, is rather different. Some researchers now believe that the origins of the Mississippi Valley seis-micity in this place have a rather more dramatic cause. They believe that here the North American Plate may be trying to split itself into two new plates along the line of major seismicity. Explorations of the deep underneath of the region suggest that material might be welling up out of the mantle well below the earthquakes’ epicenters. And, while this does not detract from the analogy of the pie—piecrusts can split into two if suitably stressed—it does have awesome implications for the region. Missouri would then become in a geological sense rather like Iceland, but covered with soybeans, cotton, and corn: It would turn into a place peppered with slow-moving volcanoes and hot springs and cliffs that move apart from one another. And there would in time be two Americas, drifting apart at the rate that fingernails grow, and with an as yet unnamed ocean in between, and somewhere drowned deep below, New Madrid, in what was once called Missouri.

  No one can tell how much stress has built up in the millions of years since the Appalachians and the Rockies and all the other plate-edge mountains were created; nor can anyone say how long it will be before the middle of the continent achieves the kind of stasis that rules out all future earthquakes, forever. The only way one can make any attempt at rationally planning for earthquakes in places like this, where, generally speaking, earthquakes do not happen, is to look very closely at those places where they have, albeit very infrequently, taken place. By doing this, one has a faint hope of imagining what could take place at some infuriatingly unspecifiable time in the future: It is only by looking at what has occurred in years gone by that one can imagine what might yet occur.

  The most venerable of the guiding mantras of geology is that the present is the key to the past. But in the very different world of seismicity it seems more prudent to suggest the converse, that the past is the key to the present. And, of course, to the future.